Transifex Interview

Design

• The Client is responsible for connecting and receiving the corresponding data from an endpoint.
• The Parser is responsible for the parsing of this data.
• The Transifex Client is responsible for communicating with the transifex api.
• The Pipeline class is responsible for creating/updating a resource.

graph TD;
Client-->Parser;
Parser-->TransifexClient;

Loading

Each pipeline is executed asynchronously and using the run_pipelines class method multiple pipelines are executed in parallel.

Any combination of a client and a parser can be used to create a pipeline. Thus, the design is modular and allows the possibility to change the parser and add other clients.

Implementation

The application uses the FastAPI backend framework and the uvicorn asgi server. In order to run it:

uvicorn main:app --host 0.0.0.0 --port 8000

The application has an endpoint for creating or updating a resource. The client has to provide a list of categories and the application creates the pipelines and executes them in parallel. A cli tool is developed under the cli folder.

python client.py

Testing, Docker and CI/CD pipeline

The application is dockerized.

docker build -t image_name .
docker run -d --name container_name -p 8000:8000 image_name

Using github actions the application is tested and deployed to a private dockerhub registry. For unnitests the pytest library is used.

pytest

More Info

• I didn't use environment variables for my account to keep the solution simple.
• Only the parser class has unittests, the test_app file has a test which tests the endpoint of the application without mocking. Mocking must be implement for the external apis.